


**********Main Text**********



****Load Study_1_Data.dta



****Figure 1

**Difference in Means Between Automated and Non-Automated Treatment
regress CredibilityBetween NonAutomated_Treatment Automated_Treatment, noconst robust 

lincom NonAutomated_Treatment - Automated_Treatment

**Generate Figure 1
forvalues i=0/1 {
capture drop x`i' d`i'
kdensity CredibilityBetween if Automated_Treatment== `i', generate(x`i'  d`i') kernel(gaussian)
}

gen zero = 0

twoway rarea d0 zero x0 if x0>=0 & x0<=100, color("15 178 170%40") ///
   ||  rarea d1 zero x1 if x1>=0 & x1<=100, color("red%15") ///
   xtitle("Threat Credibility: Probability Russia Will Use Nuclear Weapons") ///
   ytitle("Density") ///
   xla(0(10)100, nogrid) ylab(0(0.005).015) ///
   xline(64.15, lcolor("red%15")) xline(51.81, lcolor("15 178 170%40"))


   
****Figure 2
 
**Difference in Means Between Automated and Non-Automated Treatment
regress Effectiveness NonAutomated_Treatment Automated_Treatment, noconst robust 

lincom NonAutomated_Treatment - Automated_Treatment

**Figure 2 Values 

*Non-Automated Treatment (1 = Storngly Support; 7 = Strongly Oppose
tab Effectiveness if NonAutomated_Treatment==1
 
*Automated Treatment (1 = Storngly Support; 7 = Strongly Oppose
tab Effectiveness if Automated_Treatment==1
 
**Generate Figure 2
 
*Use Figure2_Data.csv
separate percentage, by(condition_number) veryshortlabel

graph bar percentage1 percentage2, ////
graphregion(color(white)) ///
over(condition_number, sort(order) label(labsize(small))) ///
over(dv, relabel(1 "" 2 "" 3 "" 4 "Threat Effectiveness: Opposition to Violating Russia's Red Line" 5 "" 6 "" 7 "")) ///
blabel(bar, position(outside) format(%9.1f) color(black)) ///
ytitle("Percentage") ///
bar(1, bcolor(blue)) bar(2, bcolor(red)) ///
ylabel(, angle(horizontal)) ///
ylabel(0(5)25) yscale(range(0(5)25)) ///
nofill ///
intensity(*1) 



****Reload Study_1_Data.csv



****Absolute Support for Violating Putin's Red Line 
summarize EffectivenessTreatmentBinary if Automated_Treatment==1

summarize EffectivenessTreatmentBinary if NonAutomated_Treatment==1



****Figure 3

**Difference in Means Between Automated and Non-Automated Treatment
regress AccidentBetween NonAutomated_Treatment Automated_Treatment, noconst robust 

lincom NonAutomated_Treatment - Automated_Treatment

**Generate Figure 3
forvalues i=0/1 {
capture drop x`i' d`i'
kdensity AccidentBetween if Automated_Treatment== `i', generate(x`i'  d`i') kernel(gaussian)
}

drop zero
gen zero = 0

twoway rarea d0 zero x0 if x0>=0 & x0<=100, color("15 178 170%40") ///
   ||  rarea d1 zero x1 if x1>=0 & x1<=100, color("red%15") ///
   xtitle("Probability of Russia Accidently Using Nuclear Weapons") ///
   ytitle("Density") ///
   xla(0(10)100, nogrid)  ///
   xline(39.08, lcolor("red%15")) xline(25.78, lcolor("15 178 170%40"))




**********Appendix**********



****Table A.1
regress CredibilityBetween Automated_Treatment, robust 

regress Effectiveness Automated_Treatment, robust 

regress CredibilityTreatmentBinary Automated_Treatment, robust 

regress EffectivenessBinary Automated_Treatment, robust 

regress CredibilityBetween Automated_Treatment if Attention==1, robust 

regress Effectiveness Automated_Treatment if Attention==1, robust 

regress CredibilityBetween Automated_Treatment Labour ideology hawkishness_index natosupport milaidukraine aiknowledge ir_attention education age income region Woman White, robust 

regress Effectiveness Automated_Treatment Labour ideology hawkishness_index natosupport milaidukraine aiknowledge ir_attention education age income region Woman White, robust 



****Table A.3
regress CredibilityBetween i.Automated_Treatment##c.ir_attention Labour ideology hawkishness_index natosupport milaidukraine aiknowledge education age income region Woman White, robust 

regress CredibilityBetween i.Automated_Treatment##i.IR_Attention_Binary Labour ideology hawkishness_index natosupport milaidukraine aiknowledge education age income region Woman White, robust 
   
regress Effectiveness i.Automated_Treatment##c.ir_attention Labour ideology hawkishness_index natosupport milaidukraine aiknowledge education age income region Woman White, robust 

regress Effectiveness i.Automated_Treatment##i.IR_Attention_Binary Labour ideology hawkishness_index natosupport milaidukraine aiknowledge education age income region Woman White, robust 
   


****Table A.4
regress CredibilityBetween i.Automated_Treatment##i.Labour ideology hawkishness_index natosupport milaidukraine aiknowledge ir_attention education age income region Woman White, robust 

regress Effectiveness i.Automated_Treatment##i.Labour ideology hawkishness_index natosupport milaidukraine aiknowledge ir_attention education age income region Woman White, robust 

regress CredibilityBetween i.Automated_Treatment##c.hawkishness_index Labour ideology natosupport milaidukraine aiknowledge ir_attention education age income region Woman White, robust 

regress Effectiveness i.Automated_Treatment##c.hawkishness_index Labour ideology natosupport milaidukraine aiknowledge ir_attention education age income region Woman White, robust 
 
regress CredibilityBetween i.Automated_Treatment##c.natosupport Labour ideology hawkishness_index milaidukraine aiknowledge ir_attention education age income region Woman White, robust 

regress Effectiveness i.Automated_Treatment##c.natosupport Labour ideology hawkishness_index milaidukraine aiknowledge ir_attention education age income region Woman White, robust 

regress CredibilityBetween i.Automated_Treatment##c.milaidukraine Labour ideology hawkishness_index natosupport aiknowledge ir_attention education age income region Woman White, robust 

regress Effectiveness i.Automated_Treatment##c.milaidukraine Labour ideology hawkishness_index natosupport aiknowledge ir_attention education age income region Woman White, robust 

regress CredibilityBetween i.Automated_Treatment##c.aiknowledge Labour ideology hawkishness_index natosupport milaidukraine ir_attention education age income region Woman White, robust 

regress Effectiveness i.Automated_Treatment##c.aiknowledge Labour ideology hawkishness_index natosupport milaidukraine ir_attention education age income region Woman White, robust 
  


****Figure A.3
eststo a: regress CredibilityWithin NonAutomated_Treatment Automated_Treatment, noconst robust 

eststo b: margins, expression(_b[Automated_Treatment] - _b[NonAutomated_Treatment ]) post

eststo c: regress EffectivenessBinaryWithin NonAutomated_Treatment Automated_Treatment, noconst robust

eststo d: margins, expression(_b[Automated_Treatment] - _b[NonAutomated_Treatment ]) post

coefplot a b, bylabel(Threat Credibility) || c d, bylabel(Threat Effectiveness) ///
||, horizontal xline(0) byopts(compact cols(1)) ci(95) xlabel(0(5)30, format(%9.0f)) /// 
xscale(range(0(5)30)) xtitle(" " "Impact of Nuclear Threat Treatments" "Relative to the Baseline (Percentage Points)") ///
xline(0, lcolor("219 68 55")) mcolor(black) msize(medium) msymbol(circle) mlabel format(%9.1f) /// 
mlabposition(12) mlabgap(1.5) mlabcolor(black) graphregion(color(white) lcolor(white) /// 
ilcolor(white)) plotregion(fcolor(white) lcolor(white) ilcolor(white)) ciopts(recast(rcap) /// 
color(black)) offset(0) legend(off) ///
coeflabels(NonAutomated_Treatment = `""Non-Automated" "Nuclear Threat""' Automated_Treatment = `""Automated" "Nuclear Threat""' ) 

eststo clear 



****Figure A.4

**Threat Perceptions
eststo a: regress ThreatPerceptionWithin NonAutomated_Treatment Automated_Treatment, noconst robust 

**No Difference Between Automated and Non-Automated Treatments (referenced in the text)
margins, expression(_b[Automated_Treatment] - _b[NonAutomated_Treatment ]) post

**Nuclear Disarmament 
eststo b: regress NukeDisarmWithin NonAutomated_Treatment Automated_Treatment, noconst robust

**No Difference Between Automated and Non-Automated Treatments (referenced in the text)
margins, expression(_b[Automated_Treatment] - _b[NonAutomated_Treatment ]) post

**Military Spending
eststo c: regress MilSpendingWithin NonAutomated_Treatment Automated_Treatment, noconst robust

**No Difference Between Automated and Non-Automated Treatments (referenced in the text)
margins, expression(_b[Automated_Treatment] - _b[NonAutomated_Treatment ]) post

**Generate Figure A.4
coefplot a, bylabel(Threat Perception Towards Russia) || b, bylabel(Support for Nuclear Disarmament) || c, bylabel(Support for Increasing Military Spending) ///
||, horizontal xline(0) byopts(compact cols(1)) ci(95) xlabel(-.5(.1).5, format(%9.1f)) /// 
xscale(range(-.5(.1).5)) xtitle(" " "Impact of Nuclear Threat Treatments" "Relative to the Baseline (5-Point Scale)") ///
xline(0, lcolor("219 68 55")) mcolor(black) msize(medium) msymbol(circle) mlabel format(%9.2f) /// 
mlabposition(12) mlabgap(1.5) mlabcolor(black) graphregion(color(white) lcolor(white) /// 
ilcolor(white)) plotregion(fcolor(white) lcolor(white) ilcolor(white)) ciopts(recast(rcap) /// 
color(black)) offset(0) legend(off) ///
coeflabels(NonAutomated_Treatment = `""Non-Automated" "Nuclear Threat""' Automated_Treatment = `""Automated" "Nuclear Threat""' ) 

eststo clear 






